The combined model with both PEV and student-level proportional decay typically fit better than either feature alone¶

Claim · All claims
Evidence · 1 study · design r2 · q2
Currency · newest study 2022 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? The RKC,SEPFA model including both features typically achieved the lowest cross-validated RMSE, implying the two features do not have identical effects. → Eglington 2022

Evidence¶

Eglington 2022¶

Eglington, L.G., Pavlik, P.I. (2022). How to Optimize Student Learning Using Student Models That Adapt Rapidly to Individual Differences. International Journal of Artificial Intelligence in Education. https://doi.org/10.1007/s40593-022-00296-0

q2 · i? · design · r2

Comparison of the five model variants' cross-validated RMSE across six datasets (Table 2); the rightmost column (RKC,SEPFA) shows the lowest RMSE in most datasets, e.g., .3954 (Andes), .4554 (Assistments), .4351 (MHE), .4006 (Cloze). No effect size printed.

"The effect on RMSE was similar, although the model with both features included typically outperformed either separately (rightmost column, Table 2) which implies they do n ot have the same effects."

Discussion¶

Related Claims¶

• Learner models with individual-difference features (PEV, student-level proportional correctness) improve cross-validated fit on six educational datasets — a broader claim this one bears on
• No single learner model was best across the six datasets, justifying a broad multi-model approach — related
• Adding PEV or student-level proportional correctness significantly reduces the absolute value of student-level signed error in all six datasets — related